UoA Maths 108 Written by Padraic Bartlett

Maths 108: Exam Review

Lectures 1-32 UoA 2021

Hello! I’ve had a number of students write in and ask for more resources for the exam,
and in particular for more examples to study before taking the exam. This document is
meant to serve as such a resource! This document is a list of what we’ve covered in this
class, along with examples to look at.
This document is not meant to serve as a replacement for your coursebook; instead, it’s
meant to give you a set of examples and summarize what we’ve been doing in the course
thus far! A good way to approach this document is to read through it, and each time you
come across something confusing use your resources (office hours, email, talking to friends,
Piazza, the coursebook, the textbooks) to review the related concepts and figure out what’s
going on.

1 Sets and Relations [Lecture 1]

We’re hoping that you can do the following tasks from this lecture:

• Describe sets using the right notation and symbols.

• Combine sets using set operations.
• Understand the definition of a function.

1.1 Sets: Fundamentals

A set is just a collection of things, which we surround with curly braces. For example,
{Bulbasaur, Charmander, Squirtle} is a set! We call the things inside a set the elements of
the set, and use the symbol ∈ to say that an element is in a set: so, for example, we could
write 2 ∈ {1, 2, 3}. To describe sets with infinitely many elements, we often describe these
sets by giving a rule that describes all of the elements of the set. For example,

{x | x = 2k + 1, k ∈ Z}

means “The set of all values of x such that x = 2k + 1, where k is an integer;” in other
words, odd numbers!
We have many famous kinds of sets: the empty set ∅ = {}, the positive whole numbers
N, the integers Z, the rational numbers Q, and real numbers R. We also had intervals of
various kinds:
=(-∞,-1] =(0,1) =[1.5,2.5) =(3,+∞)

- ∞ -2 -1 0 1 2 3 +∞
Given two sets A, B, we can combine them in several ways. The intersection of A and B,
written A ∩ B, is the set of all elements that are in both A and B; for example, {1, 2, 3} ∩
[0, 2) = 1. The union of A and B, written A ∪ B, is the set of all elements in either A or
B or both; for example, (−∞, 2) ∪ [1, 3] = (−∞, 3]. Finally, A set-minus B, written A \ B
or A − B, is the set of all elements in A that are not in B; for example, [3, 4] \ Z = (3, 4).

1
1.2 Relations and Functions
A function f : A → B consists of three things:
• A set A, called the domain. We think of A as the set of all possible “inputs” to the
function f .
• A set B, called the codomain. We think of B as containing all of the possible
“outputs” that the function f could have.
• A rule f , that assigns elements in A to elements in B. To be a function, for every a
in A there must be exactly one b ∈ B such that f (a) = b; you can’t have a value of a
where f (a) is undefined, or a value of a that we try to send to two different values in
B!
For example, consider the following three objects:
√
1. f : R → R, f (x) = x
√
2. g : [0, ∞) → R, f (x) = x
3. h : [−1, 1] → R, f (x) = y if and only if x2 + y 2 = 1.
The first
√ object is not a function, because it’s not defined on every element of its domain:
f (−1) = −1 is not an element of R, for example!
The second object is a function: for any nonnegative number x, there is exactly one
√
value of y in R equal to x; so every element in the domain has exactly one corresponding
element in the codomain.
The third object is not a function. Perhaps the easiest way to see this is to graph it:
that is, let’s draw all of the points (x, y) where x2 + y 2 = 1! As you’ve seen before, this
forms a circle:
= graph
of h

= vertical
line
test

For every x ∈ [−1, 1], there is a corresponding √value of y such that x2 + y 2 = 1; namely,
2 2
if we solve for y, we get that y = 1 − x ⇒ y = ± 1 − x2 , which exists for any x ∈ [−1, 1].
The issue, however, is there there are *two* different values of y for our value of x! A
function is supposed to have exactly one output, but we have two here; as a result, this
object is not a function.
A nice way to visualize whether or not something is a function, in the case that you’re
dealing with things you can graph, is the vertical line test. If you draw a graph of some rule,
look at vertical lines of the form x = a for any value a in your domain A. If that vertical
line hits your graph more than once, you’re not a function: that value of x has multiple
values of y that correspond to it! Conversely, if that vertical line does not hit your graph
at all, you’re also not a function: there is no output that corresponds to that input a from
the domain A!

2 Functions [Lectures 2-3]

After finishing this lecture, we’re hoping that you can do the following tasks:
• Visually identify the graphs of several basic functions, and know various key values
and properties of these functions.
• Know how to combine functions with arithmetic and composition.

2
2.1 Functions: Fundamentals
When people think about functions, they often just think about the rule that defines a
x2 + 1
function: that is, we’ll want to regard something like f (x) = 2 as a function, even
x −1
though we haven’t specified its domain or codomain yet!
In cases like this, it is useful to be able to find out what the “largest” domain we could
assign this function that would make sense. For instance, the expression f above is defined
as long as x 6= ±1; so we could say that it is a function with domain R \ {±1}, and
codomain R.
A common question that we’ll ask you in Maths 108 is the following: given an expression
f , can you find a set of values on which f is a function? As an example, the hardest problem
on your mid-semester test (74% of students got this wrong) was the following:

1
3. Suppose that the rule f is given by the formula f (x) = tan(x) .
π 3π


Consider the three following sets: (π, 2π), 2 , 2 , (0, π).
On how many of these sets is f a function?

(a) 0 (b) 1 (c) 2 (d) 3

The answer here is that f is not defined on any of these sets. This is because of the
following logic:

• We cannot divide by zero.

• As a result, because cos(x) is zero whenever x is equal to (2k+1)π
2 for some integer k,
sin(x) (2k+1)π
tan(x) = cos(x) does not exist whenever x = 2 .
• tan(x) is also equal to 0 whenever sin(x) = 0, which happens when x = kπ for some
integer k.
1
• As a result, tan(x) does not exist whenever x = kπ for some integer k (as it would
make us divide by 0) or x = (2k+1)π
2 (as this would mean we’re dividing by something
that’s undefined. This eliminates all of the sets shown above.

Note that when you’re working with expressions in this course you need to be very
1
careful with simplifying! Many people turned tan(x) into cot(x) = cos(x)
sin(x) , which is true
whenever x 6= (2k+1)π
2 ; however, when x = (2k+1)π2
1
, cot(x) is defined and tan(x) is not!
1 1
This is like how 1/x and x are equal if x 6= 0, but when x = 0 the expression 1/x is
undefined (and so these expressions are regarded as being distinct.)
√
The functions sin(x), cos(x), tan(x), ex , ln(x), |x|, x, x1 and polynomials are all functions
that we use frequently in Maths 108. We draw them here:

Graph: y = poly Graph: y = |x| Graph: y = x1

Nat. dom: R Nat. dom: R Nat. dom: R \ {0}

3
 √
Graph: y = ex Graph: y = ln(x) Graph: y= x
Nat. dom: R Nat. dom: (0, ∞) Nat. dom: √ ∞)
[0,
Key values: e0 = 1 Key values: ln(1) = 0 Key values: 0=0

Graph: y = sin(x) Graph: y = cos(x) Graph: y = tan(x)

Nat. dom: R Nat. dom: R Nat. dom: R \ {kπ + π2 | k ∈ Z}.
Key values: sin(0) = 0, Key values: cos(0) = 1,√ Key values: tan(0) = 0,
sin( π6 ) = 1√
2 , cos( π6 ) = √23 , tan( π4 ) = 1,
π 2
sin( 4 ) = √2 , cos( π4 ) = 22 , tan( π2 ) DNE
sin( π3 ) = 23 , cos( π3 ) = 12 ,
sin( π2 ) = 1 cos( π2 ) = 0
These functions have several useful properties. For all x ∈ R, we know that

sin(x)
• sin(x + π) = − sin(x) • cos(−x) = cos(x) • tan(x) = cos(x)
• cos(x + π) = − cos(x) • sin(x + 2π) = sin(x) • sin2 (x) + cos2 (x) = 1.
• sin(−x) = − sin(x) • cos(x + 2π) = cos(x) • ln(ex ) = x.

Understanding how sin and cos can be derived from the unit circle can be a much more
efficient way to remember all of the trigonometric identites and key values above.

2.2 Combining Functions

Given any of these building block functions, we can combine them in several ways. Given
two functions f (x), g(x), we can add, subtract, multiply, and divide these functions by each
other to create new functions. We can also compose these functions with each other, to
create a new function (f ◦ g)(x) = f (g(x)); for example, if f (x) = x2 − 1, g(x) = x + 1 we
would have (f ◦ g)(x) = f (g(x)) = f (x2 + 1) = (x2 + 1)2 − 1 = x2 + 2x.

3 Limits [Lectures 4-6]

After finishing this lecture, we’re hoping that you can do the following tasks:

• Evaluate limits of various kinds.

• Understand what the vertical and horizontal asymptotes of a function are, and how
to draw these objects.
• Determine whether a given function is continuous.

4
3.1 Kinds of Limits
Given a function f (x), we’ve studied several different kinds of limits in Maths 108:

• lim f (x) = L. Roughly speaking, this means the following: as we plug in any values
x→a
of x that get closer and closer to a, we get outputs f (x) that get closer and closer to
L.
• lim f (x) = L and lim f (x) = L. This is the same idea as the above (plug in any
x→a− x→a+
value of x close to a, get output f (x) close to L), except that for the a− limit we add
the restriction that x < a, and for the a+ limit we add the restriction that x > a. We
call these the limits as we approach a from the left or the right, respectively.
• lim f (x) = L and lim f (x) = L. Same idea again, except in the −∞ case we’re
x→−∞ x→+∞
plugging in arbitrarily large and negative values of x, while in the +∞ case we’re
plugging in arbitrarily large and positive values of x.

Notice that limits don’t care about what happens when x gets to a — in the definitions
above, we never plug a into f (x)! Instead, we always plug in values of x that get closer and
closer to a. This lets us say things like
x
lim = 1,
x→0 x

x
even though the function x is undefined at zero.

3.2 Limit Techniques

We have several ideas that are useful for evaluating and working with limits:

3.2.1 Limit Idea 1: Plugging In Values

Probably the simplest thing you can do, when given a limit, is just physically plug in num-
1
bers and figure out where the function is going. For instance, to find the limit lim , we
x→0 x
+
could just plug in various values of x that are close to but greater than zero, and see what
happens:
x f(x)
0.1 10
0.01 100
0.001 1000
0.0001 10000
It certainly looks like f (x) is growing arbitrarily large, so we could quite reasonably believe
1
that lim = +∞.
x→0+ x
This isn’t a bad technique, but it has its issues (which is why we have other ways to
find limits!) One of which is that this method is fairly prone to human error. If you have a
x3 − 3x2 + 3x
limit like lim and you’re trying to plug in something like x = −0.001 into
x→0− x2 − x
that fraction, it’s going to be really easy to drop a negative sign somewhere.
Another is that it can be pretty hard to tell whether or not you’re actually plugging in
enough values to figure out the pattern. For instance, if someone asked you for the limit
lim tan(x), and you plugged in x = π, 10π, 100π, 1000π, it would look like your limit was
x→∞
going to 0. This, however, is because we were silly and kept plugging in multiples of π; if

5
we plugged in other large numbers we’d see that actually lim tan(x) DNE, as looking at
x→∞
the graph of tan(x) from earlier demonstrates.

3.2.2 Limit Idea 2: Continuity

We say a function f (x) is continuous at x = a if lim f (x) = f (a): in other words, if the
x→a
value the function takes on at x = a is the value it looks like the function is going to take
as we plug in values of x close to a! Visually, this means that our function doesn’t have a
“jump” or “break” or goes off to infinity at x = a, and that it’s defined at x = a.
√
Conveniently, all of the building block functions sin(x), cos(x), tan(x), ex , ln(x), |x|, x, x1 ,
polynomials are all continuous where they are defined, as is everything we can make by com-
bining these functions with composition or arithmetic where they are defined! So, if someone
1
were to ask for the limit lim , you could just say “this function is continuous on R \ {0},
x→2 x
so it’s just 21 because that’s what you get when you plug in x = 2 into x1 .”
When we’re looking at limits of functions on values at which they’re not defined, or
1
when we’re working with piecewise-defined functions, things can get weird: lim DNE, for
x→0 x
1
example, while lim 2 + ∞. We use our other techniques to figure out what’s going on in
x→0 x
these cases.

3.2.3 Limit Idea 3: Showing a Limit DNE

Often, we want to show that a given limit does not exist. The following idea is useful when
doing this:

The limit lim f (x) = L exists precisely when lim f (x) = L , lim f (x) = L both
x→a x→a− x→a+
exist and are equal to each other.

1 1
So, for instance, because the limit lim = +∞ and the limit lim = −∞, the
x→0+ x x→0 x
−
1
overall limit lim = DN E.
x→0 x
This is not the only way to show that a limit does not exist; we saw earlier that
lim tan(x) DNE by looking visually at its graph! But it is often a useful technique.
x→∞

3.2.4 Limit Idea 4: Simplifying Fractions

f (x)
In the special case where your limit has the form lim where lim f (x) = 0 and
x→a g(x) x→a
lim g(x) = 0, a useful technique you can try is simpliflication! Basically, take your
x→a
f (x)
fraction , and try to simplify it by factoring the top and bottom and canceling terms.
g(x)
This often gives you a new function where you no longer have the top and bottom going to
zero, which is often much easier to work with.
x2 − x
So, for example, if we had the limit lim , we could factor an x out of the top and
x→0 x
bottom to get the simplified limit lim x − 1. This is a limit of a continuous function, so we
x→0
can just plug in x = 0 to get that the limit is 0 − 1 = −1.

6
 x2 − x
For a second example, if we had the limit lim , we could again factor an x out of
x→0 x2
x−1
the top and bottom to get the simplified limit lim .
x→0 x
If we were to plug in values of x close to 0, we can see that for tiny x < 0 it looks like
our function is going to +∞, and for tiny x > 0 it looks like our function is going to −∞:
x f(x) x f(x)
-0.1 11 0.1 -9
-0.01 101 0.01 99
-0.001 1001 0.001 999
So it seems likely that this limit does not exist, as the limits from the left and the right are
not equal.

3.2.5 Limit Idea 5: Break It Down

Our last limit idea is the following: given a limit lim f (x), we can often figure out what’s
x→a
going on with it by breaking our functions down into small pieces, looking at what those
individual pieces do as x goes to a, and then slowly “zooming back out” to see what the
whole function does.
This is an idea best illustrated with an example. Consider the limit
ex
lim
x→0 1 + ln(|x|)

Understanding this function all at once is hard! But, notice that for very small values
of x we know that
• |x| is a very small positive number, therefore
• ln(|x|) is defined and a very large negative number, therefore
• 1 + ln(|x|) is also a very large negative number.
So we know what the denominator is doing.
Conversely, we have limx→0 ex = e0 = 1, because ex is continuous. So, for very small
values of x near zero, we have
ex ∼ something very close to 1 ∼
= = 0.
1 + ln(|x|) something very large and negative
In other words, our limit is 0.

3.3 Limit Applications: Drawing Functions

Given a function f (x), we say that x = a is a vertical asymptote if the limit as we
approach a from either the right or the left is either +∞ or −∞. Similarly, we say that
y = L is a horizontal asymptote for any real number L if the limit as x approaches either
positive infinity or negative infinity is L.
Often, knowing a function’s vertical and horizontal asymptotes can help us create a quick
2 −4
sketch of what that function looks like. For instance, take the function 2xx2 −9
. As x goes
to either positive or negative infinity, the −4 in the numerator is going to be insignificant
in comparison to the 2x2 term, and similarly the −9 isn’t relevant in comparison to the x2
in the denominator. So for large positive or negative values of x, this function is basically
2
equal to 2x
x2
= 2; so we have

2x2 − 4 2x2 − 4
lim = 2, lim = 2,
x→+∞ x2 − 9 x→−∞ x2 − 9

7
and therefore that 2 is our only horizontal asymptote.
For vertical asymptotes, the only places where our function isn’t continuous are where
the denominator is zero, so these are the only places where our function might go to ±∞.
This happens when x2 − 9 = 0, i.e. at x = ±3; so we want to understand the limits at these
two values.
If we look at −3 in particular, we can see that the numerator 2x2 −4 is about 2·9−4 = 14.
At the same time, the denominator x2 −9 is going to zero: in particular, if we plug in values of
x close to but less than −3, we get that x2 is something slightly larger than 9, and therefore
14
that x2 − 9 is something tiny and positive. Therefore, the ratio tiny, positive is going to +∞,
2x2 − 4
and so the limit lim = +∞.
x→−3− x2 − 9
Similarly, if we plug in values of x close to but greater than −3, x2 is slightly smaller
14
than 9, so x2 − 9 is a tiny negative number. Therefore, the ratio tiny, negative is going to
2
2x − 4
−∞, and so the limit lim = −∞.
x→−3+ x2 − 9
In particular, x = −3 is a vertical asymptote! If you use the same reasoning at x = 3,
2x2 − 4 2x2 − 4
you’ll see that lim 2 = +∞, lim 2 = −∞.
x→3+ x − 9 x→3− x − 9
Now, simply draw in these lines y = 2, x = −3, x = 3 and sketch a curve that has the
limits we just found. You’ll get something like what we have below, which is the graph of
our function!

4 Vectors and their Operations [Lectures 7-9]

After finishing this lecture, we’re hoping that you can do the following tasks:
• Perform various vector operations, like addition, scalar multiplication, measuring
length, and finding the dot product.
• Identify when a vector is a linear combination of other vectors.
• Use the dot product to measure angles, and in particular tell when two vectors are
orthogonal.

4.1 Vector Addition and Scalar Multiplication

A vector is any element of Rn . So, for example, (1, 2) ∈ R2 and (0, 3, 4) ∈ R3 are both
vectors. Given any two vectors with the same number of coordinates, we can add these
vectors by summing them coordinate-by-coordinate: so, for example, we have

(1, 2, 4, 3) + (0, 1, 5, 5) = (1, 3, 9, 8).

A scalar is another word for a real number. Given any scalar a ∈ R and any vector v ∈ Rn ,
we can find the scalar multiple av by taking v and multiplying each of its coordinates by
a: so, for example,

3(1, −2, 0) = (3, −6, 0).

8
We say that two vectors are parallel if one is a scalar multiple of the other.
Geometrically, these operations are nice to visualize:
y

2u= (2,6) y

u=(2,3)

u=(1,3)
u+v = (5,1)
x
x
-1u=(1,3)

v=(3-2)

Given any set of vectors v1 , v2 , . . . vk ∈ Rn , a linear combination of these vectors is

anything we can make by adding together scalar multiples of these vectors. So, for example,
(1, 2, 3) is a linear combination of the two vectors v = (2, 3, 4), w = (3, 4, 5), because we can
write

(1, 2, 3) = 2(2, 3, 4) − (3, 4, 5).

Conversely, (1, 0, 0) is not a linear combination of the two vectors (0, 2, 3), (0, 3, 2), because
for any scalars a, b the linear combination

a(0, 2, 3) + b(0, 3, 2) = (0, 2a + 3b, 3a + 2b)

cannot be equal to (1, 0, 0), as they will never have the same first coordinate.

4.2 Length
As well, given any vector in Rn , we can find its length by using Pythagoras’s theorem. To
be precise, given any vector v = (v1 , v2 , . . . vn ), the length of v is given by the expression
q
v12 + v22 + . . . + vn2 .

So, for example, the length of the vector (2, 6, 9) is

p √ √
22 + 62 + 92 = 4 + 36 + 81 = 121 = 11.

We say that a vector is a unit vector if it has length 1. So, for instance, ( 35 , 45 ) is a
unit vector, because
s 
3 2
 2 r
4 9 16 √
+ = + = 1 = 1.
5 5 25 25

4.3 The Dot Product

Given two vectors v = (v1 , v2 , . . . vn ), w = (w1 , w2 , . . . wn ), we can form their dot product
v · w as follows:

v · w = v 1 w1 + v 2 w2 + . . . + v n wn .

9
So, for example, we have

(1, 2, 3) · (2, 3, 1) = 1 · 2 + 2 · 3 + 3 · 1 = 11, (1, 0, 0) · (0, 1, 0) = 1 · 0 + 0 · 1 + 0 · 0 = 0.

Given any two vectors u, v, let θ denote the angle between u, v. We can measure θ
using the dot product:
 
−1 u·v
θ = cos
||u||||v||
For example, if u = (1, 1, 1, 1) and v = (3, 1, 1, 1), the angle between these vectors is
   
−1 (1, 1, 1, 1) · (3, 1, 1, 1) −1 3+1+1+1
θ = cos = cos √ √
||(1, 1, 1, 1)|| · ||(3, 1, 1, 1)|| 1+1+1+1+ 9+1+1+1
    √ !
6 6 3 π
= cos−1 √ = cos−1 √ = cos−1 = .
2 12 4 3 2 6

We say that two vectors are orthogonal if their dot product is zero. Geometrically,
this means that the angle between these two vectors is π/2. For example, the two vectors
(2, 3) and (3, −2) drawn earlier are orthogonal, because (2, 3) · (3, −2) = 6 − 6 = 0. In the
picture, you can see that these two vectors are meeting at a right angle.

5 Lines and Plane: A Field Guide [Lectures 10-12]

The main things we’re hoping you can do after finishing these lectures are the following:
• Describe a line through two points, through a point parallel to a vector, with para-
metric equations, or with general equations.
• Describe a plane through three points, through a point parallel to a pair of vectors,
with parametric equations, with a general equation, or with a point-normal equation.
• Convert one way of describing a line/plane into another way of describing the same
line/plane.
• Given any (point/line/plane) and any other (point/line/plane), find and describe the
intersection of these two objects.

5.1 Lines
5.1.1 A Line Through Two Points
Given two points P 6= Q ∈ Rn , the line through P and Q has equation given by

x = P + t(Q − P ), t∈R

Example. The line through the two points (1, 0), (0, 1) in R2 has equation

(x, y) = (1, 0) + t((0, 1) − (1, 0)) = (1, 0) + t(−1, 1), t∈R

We can draw this line by plugging in various values of t into our equation:
y
t (x,y)
-1 (2,-1)
x 0 (1,0)
1 (0,1)
2 (-1,2)

10
5.1.2 A Line Through a Point, Parallel to a Vector
Given a point P and vector v 6= 0 ∈ Rn , the line through P parallel to v has equation
given by

x = P + tv, t∈R

Example. The line through (1, 1, 1) parallel to the vector (0, 1, −1) in R3 has equation

(x, y, z) = (1, 1, 1) + t(0, 1, −1), t ∈ R.

We draw this line here in purple, along with the vector (0, 1, −1) that it is parallel to in gold:
z
t (x,y,z)
-1 (1,0,2)
y 0 (1,1,1)
1 (1,2,0)
x
2 (1,3,-1)

5.1.3 Parametric Equations for a Line

A set of parametric equations for a line in Rn is a set of expressions of the form a + bt
where a, b are constants, where we have one such expression for each coordinate.

Example. We saw earlier that line through the point (1, 1, 1) parallel to the vector (0, 1, −1)
in R3 has equation

(x, y, z) = (1, 1, 1) + t(0, 1, −1), t∈R

We can express this as a set of parametric equations by writing out what the equation above
tells us x, y, z are equal to:

x = 1,
y = 1 + t,
z = 1 − t, t ∈ R.

5.1.4 General Equations for a Line

The general equation for a line in R2 is probably the first way you saw lines described. In
R2 , a general equation for a line in R2 is any equation of the form

ax + by = c.

In R3 , things get weirder. A general equation for a line in R3 is any pair of equations of
the form

ax + by + cz = d,
ex + f y + gz = h,

as long as that the two planes described by these two equations are not parallel.

11
Example. We can graph the line with general equations

x + y + z = 1,
x − y + 2z = −1

by plugging in values of x into both equations and using this to solve for corresponding
values of y and z. For instance, if x = 0 we have y + z = 1, −y + 2z = −1, which when
added tells us that 3z = 0, i.e. z = 0, which forces y = 1, and gives us the point (0, 1, 0).
Doing this several times generates the graph below:
z

x y z
0 1 0
y
3 0 -2
-3 2 2
x

5.2 Planes
5.2.1 A Plane Through Three Points
Given three points P, Q, R ∈ Rn not all contained on the same line, the plane containing
the three points P, Q, R has equation given by

x = P + s(Q − P ) + t(R − P ), s, t ∈ R

Example. The plane containing the three points (1, 1, 0), (1, 0, 1), (0, 1, 1, )in R3 has equa-
tion

(x, y, z) = (1, 1, 0) + s((1, 0, 1) − (1, 1, 0)) + t((0, 1, 1) − (1, 1, 0))

= (1, 1, 0) + s(0, −1, 1) + t(−1, 0, 1) s, t ∈ R

We can draw this plane by plugging in various values of s, t into our equation.
z

s t (x,y,z)
1 1 (0,0,2)
y
1 -1 (2,0,0)
-1 1 (0,2,0)
x

5.2.2 A Plane Through a Point, Parallel to Two Vectors

Given a point P and two nonparallel vectors v, w ∈ Rn , the plane containing P parallel to
~v , w
~ has equation given by

x = P + sv + tw, s, t ∈ R

Example. The plane containing the points (0, 0, 3) parallel to the vectors (2, 0, 0), (0, −3, 0)in
R3 has equation

(x, y, z) = (0, 0, 3) + s(2, 0, 0) + t(0, −3, 0) s, t ∈ R

12
5.2.3 Parametric Equations for a Plane
A set of parametric equations for a plane in Rn is a set of expressions of the form
a+bs+ct where a, b, c are constants, where we have one such expression for each coordinate.

Example. We saw earlier that the plane containing the three points (1, 1, 0), (1, 0, 1), (0, 1, 1)
in R3 has equation

(x, y, z) = (1, 1, 0) + s(0, −1, 1) + t(−1, 0, 1) s, t ∈ R

We can express this as a set of parametric equations by writing out what the equation above
tells us x, y, z are equal to:

x = 1 − t,
y = 1 − s,
z = s + t, s, t ∈ R.

5.2.4 General Equations for a Plane

A general equation for a plane in R3 is an equation of the form ax + by + cz = d.

Example. The equation x + y + z = 2 is a general equation for a plane. In fact, because

the three points (1, 1, 0), (0, 1, 1), (1, 0, 1) satisfy this equation, this is the same plane that
we gave as an example for the equation for a plane through three points!

5.2.5 Point-Normal Equations for a Plane

A vector n is called orthogonal to a plane P if the line x = tn intersects the plane P at a
right angle. Given a plane with general equation ax + by + cz = d, the vector (a, b, c) is a
normal vector to that plane, as is any nonzero scalar multiple of (a, b, c).
Given a point P and a vector n 6= ~0, a point-normal equation for a plane through
that point P with n as its normal vector is the following:

n · (~x − P ) = 0.

Example. The plane x+y+z = 2 from earlier has (1, 1, 1) as a normal vector. If we combine
this information with our knowledge that this plane goes through the point (1, 1, 0), we can
create a point-normal equation for this plane:

(1, 1, 1) · ((x, y, z) − (1, 1, 0)) = 0.

Why is this useful? Well: given two planes P, Q, it turns out that the angle of intersection
of these two planes is precisely the angle of intersection of their normal vectors! In other
words, if you have two planes P, Q intersecting at some angle θ, then the angle between the
normal vectors to these planes is also equal to θ.

Example. The angle of intersection θ of the plane x+y+z = 2 and the plane x+2y−3z = 4
is the same thing as the angle of intersection θ of the normal vectors (1, 1, 1) and (1, 2, −3),
which is
 
−1 (1, 1, 1) · (1, 2, −3) π
θ = cos = cos−1 (0) = .
||(1, 1, 1)||||(1, 2, −3)|| 2

13
6 Switching Between Different Equations
6.1 Lines
Parametric equations, equations through two points, and equations through a point parallel
to a vector for lines are almost identical: there’s really no difference between saying

(x, y) = (1, 0) + t((0, 1) − (1, 0)), t ∈ R,

or (x, y) = (1, 0) + t(−1, 1), t ∈ R,
or x = 1 − t, y = t, t ∈ R.

Converting between general and parametric equations, then, is really the only interesting
thing. You can do this as follows:

1. Given general equations for a line, choose a variable in your equations that’s not
constant. Set it equal to t. Now, plug this into your other equations, and solve for
the other variables in terms of t.
2. Given parametric equations for a line, choose one equation and solve for t in terms of
one of your coordinates. Use this to get rid of t in your other equations.

Example. Given the line with general equations 2x + y = 2, z = 1, start by picking a

variable that’s not constant — in this case, either x or y — and set it equal to t. For
example, we’ll pick x = t. Now, we plug this into our equations, to get 2t + y = 2, z = 1.
Solving for y gives us y = 2 − 2t, and thus the parametric equations

x=t
y = 2 − 2t
z = 1, t ∈ R.

Example. Given the line with parametric equations x = 2t − 1, y = t + 3, z = 2 − t,

we can find a set of general equation for this line. Start by picking an equation, say
y = t + 3, and solving for t; i.e. t = y − 3. Plug this into the other equations to get
x = 2(y − 3) − 1, z = 2 − (y − 3); in other words, the general equations

x − 2y = −7, z + y = 5.

6.2 Planes
Similarly, it’s very easy to switch between parametric equations, equations through three
points, and equations through a point parallel to two vectors for a plane:

(x, y, z) = (1, 0, 0) + s((0, 1, 0) − (1, 0, 0) + t((0, 0, 1) − (1, 0, 0)), s, t ∈ R,

and (x, y, z) = (1, 0, 0) + s(−1, 1, 0) + t(−1, 0, 1), s, t ∈ R,
and x = 1 − s − t, y = s, z = t, s, t ∈ R

are all basically the same thing.

As well, it’s easy to switch between general equations and point-normal equations, be-
cause the general equation of a plane tells you what the normal vector to that plane is:

(1, 1, 1) · ((x, y, z) − (2, 0, 0)) = 0 ⇔ x + y + z = 2.

So the only interesting thing, again, is switching between parametric and general equa-
tions. You can do this with the exact same methods as before:

14
 1. Given general equations for a line, choose two variables in your equations that’s not
constant. Set one equal to t, and the other equal to s. Plug this into your general
equation to express the other variable in terms of s and t.
2. Given parametric equations for a line, choose one equation and solve for t in terms of
the other variables. Plug this into your other equations to get rid of t. Now, solve for
s, and plug this into your third equation. You should now have a general equation!

Example. Given the plane with general equation x + 2y + 3z = 4, we arbitrarily decide

to set y = s and z = t. Plugging this into our equation tells us that x + 2s + 3t = 4, i.e.
x = 4−3t−2s, and therefore that our plane can be described with the parametric equations

x = 4 − 3t − 2s
y=s
z = t, s, t ∈ R.

Example. Given the plane with parametric equations x = s + t, y = s − t, z = 1 + 2s + 3t,

we take the first equation and solve for s, to get x − t = s. Plugging this into the other
equations gives us

y = (x − t) − t = x − 2t,
z = 1 + 2(x − t) + 3t = 1 + 2x + t

The third equation tells us that z −1−2x = t, which when plugged into the second equation
gives us y = x − 2(z − 1 − 2x) = 5x − 2z + 2, which when rearranged is the general equation

−5x + y + 2z = 2.

7 Intersecting Points, Lines and Planes

Given a line and a plane, or a pair of lines, or a pair of planes, we often want to find their
intersection: the set of points that these two objects share in common. Geometrically,
these are the following possibilities:
1. The intersection of two lines can either be the empty set, a point, or a line. If two
lines intersect on a line, then these two lines are the same — this is a good way to
check whether two different equations are describing the same line!
2. The intersection of a line and a plane can either be the empty set, a point, or a line.
If the intersection is a line, then the line is contained within the plane.
3. The intersection of two planes can either be the empty set, a line, or a plane. Note
that a point is impossible. If the intersection is a plane, then those two planes are
equal!
The process for finding the intersection of two objects P, Q is pretty simple, actually.
• To start, assume that you have a set of general equations for P , and a set of parametric
equations for Q. (If you don’t, use the techniques from earlier to get such equations!)
• Now, plug your parametric equations into your general equations: you’ll get equations
in t or in s, t.
• Try to solve for your variables s, t:
– If you get a contradiction: your intersection is the empty set.
– If you get exact values for s, t: your intersection is a point.
– If you have one variable that cannot be solved for: your intersection is a line.
– If you cannot solve for both s and t: you have a plane.

15
 We illustrate this process with examples:
Example. To find the intersection of the line with general equation x + 2y = 4 with the
line with parametric equations x = 3t − 1, y = t, we plug our parametric equations into our
general equation, and get
(3t − 1) + 2(t) = 4
⇒5t = 5
⇒t = 1.
This is a single value of t; therefore these two lines intersect at a point, in particular the
point we get when we plug t = 1 into the equations x = 3t − 1, y = t: the point (2, 1)!
Example. To find the intersection of the line with parametric equations x = 2 + t, y =
3 − t, z = t with the plane 2x + y − z = 0, we plug our parametric equations into our general
equation and get that
2(2 + t) + (3 − t) − (t) = 0
⇒7 = 0.
This is a contradiction: therefore, no such value of t exists, and the intersection of these
two objects is the empty set!
Example. To find the intersection of the plane with general equation 2x − y + z = 4 and
the plane with parametric equations x = 2 + s, y = 2 + t, z = 2 − 2s + t, we plug our
parametric equations into our general equation, and get
2(2 + s) − (2 + t) + (2 − 2s + t) = 4
⇒4 + 2s − 2 − t + 2 − 2s + t = 4
⇒0 = 0.
In other words, we can’t solve for either s or t! Therefore the intersection of these two
planes is a plane, and in particular this means these two planes are the same!

7.1 Points, Lines and Planes

We also sometimes want to know if a given point P is contained in a line or a plane. This is
even easier: just plug the point into your equations, and solve! For example, the line with
parametric equations x = 2 − t, y = 2 + t does not contain the point (0, 3), because plugging
in x = 0, y = 3 into these equations yields

0 =2−t
⇒ 3 = (2 − t) + (2 + t) = 4 ⇒ 3 = 4,
3 =2+t
which is impossible.
To give another example, the plane with general equation x + 2y − z = 2 contains the
point (1, 1, 1). This is because plugging in x = 1, y = 1, z = 1 into our equation yields
1 + 2 − 1 = 2, which is true, and therefore shows that (1, 1, 1) is in this plane.

8 Solving Systems of Linear Equations [Lectures 13-15]

After reading and watching these lectures, we’re hoping that you can do the following tasks:
• Take a system of linear equations and transform it into an augmented matrix.
• Use row operations to transform an augmented matrix into reduced row-echelon form.
• Use the reduced row-echelon form of an augmented matrix to find all of the solutions
to its corresponding system of linear equations.

16
8.1 Fundamentals
A linear equation in n variables x1 , . . . xn is an equation of the form a1 x1 + a2 x2 + . . . +
an xn = c, for constants a1 , . . . an , c ∈ R. For example, 2x + 3y = 4 and 3x + 2y − z = 11 are
both linear equations. A system of linear equations is just a collection of multiple linear
2x + y = 1
equations, like . Given a system of linear equations, a solution is any way
x+y =2
to replace the variables of that system with real numbers, so that all of the equations are
2x + y = 1
satisfied. For example, (x, y) = (−1, 3) is a solution to , because plugging
x+y =2
in x = −1, y = 3 into our equations satisfies both of them!
Given a system of linear equations, we often want to find all of the possible solutions
to that system of equations. In these lectures, we came up with a multi-step process for
finding these solutions! The key ingredients in this process are the following:

1. We can create the augmented matrix corresponding to a system of linear equations

by writing down the coefficients of that system of equations in a matrix. For example,
 
2x + y = 1 2 1 1
the system corresponds to the augmented matrix . We
x+y =2 1 1 2
put the coefficients of the variables on the left and put the constants on the right, and
draw a vertical line between them so that we can quickly tell the difference.
2. Given an augmented matrix, we can perform row operations on that augmented matrix
without changing the solutions to the underlying system of linear equations! There
are three row operations available to us:
• Switch two rows.
 For example,
 we could switch
 rows 1 and 2 in the augmented
2 1 1 switch R1 and R2 1 1 2
matrix above: −−−−−−−−−−−→ .
1 1 2 2 1 1
• Add amultiple of one row to another. For example, we could add −2R1
1 1 2 add −2R1 to R2 1 1 2
to R2 : −−−−−−−−−−→ .
2 1 1 0 −1 −3
• Multiply each entry ina row by a nonzero constant.  For example,
 we
1 1 2 multiply R2 by −1 1 1 2
could multiply R2 by −1: −−−−−−−−−−−→ .
0 −1 −3 0 1 3
3. The goal of these row operations is to transform this matrix into something called
reduced row-echelon form, or RREF for short. Augmented matrices in reduced
row-echelon form have the following form:
• The first nonzero entry in each row is a 1; we call these 1’s leading 1’s.
• All entries in the same column as a leading 1 are zero.
• Each leading 1 is to the left of any leading 1 below it.
 
1 1 2
For example, if we take the augmented matrix we were working on earlier
0 1 3
 
1 0 −1
and add −R2 to R1 , we get . The first nonzero entry in each row is a 1,
0 1 3
there are 0’s above and below these 1’s, and each leading 1 is the left of the leading
1 below it; so this is in reduced row-echelon form!
We call a variable a system of linear equations leading if there is a leading 1 in the
corresponding column of its augmented matrix’s reduced row-echelon form, and we
call a variable free if it is not leading.
4. Finally, once the augmented matrix is in reduced row-echelon form, we turn it back

17
  
1 0 −1 x + 0y = −1
into a set of linear equations: −→ . In this form, it’s
0 1 3 0x + y = 3
really easy to read off the solutions to our original system: we just want x = −1, y = 3!

8.2 Algorithms and an Example

So: we know what we want to do (take an augmented matrix, get its reduced row-echelon
form, interpret the answers.) To describe the “how,” I’ve created some flowcharts:

Is there a row
Yes without a leading No
We’re done!
1 that has nonzero
entries in it?

Let Ri be the
topmost row
Using row op- Make all other
without a leading
erations, make entries in column
1. Let Cj be the
the entry in (i, j) Cj zero by adding
leftmost column
a 1. Call this multiples of
containing nonzero
1 a leading 1. Ri to them.
entries other
than leading 1’s.

Figure 1: A flowchart for finding the RREF of an augmented matrix.

You have a line

of solutions.

Is there a row
No 2
 that looks like  How many free You have a plane
0 ... 0 c variables are there? of solutions.
for some c 6= 0?
3+
Yes 0

There are no There is exactly You have a d-

solutions to one solution to dimensional space of
your system of your system of solutions, where d is
linear equations. linear equations. the number of free
variables you have.

Figure 2: A flowchart for how to use RREFs to solve systems of linear equations.

These charts summarize the processes we’ve described in class for finding the RREF of
a matrix and interpreting it! To illustrate how these processes work in action, we consider
an example problem here:
Example. Consider the system of linear equations
4x+ 3y+ 2z = 1
x+ y+ z = 1
y+ 2z = 3
What are the solutions to this system of linear equations? How many solutions does this
system of linear equations have?

18
Answer. We start by converting this system of linear equations to an augmented matrix:
 
4x+ 3y+ 2z = 1 4 3 2 1
x+ y+ z = 1 −→  1 1 1 1 
y+ 2z = 3 0 1 2 3

We then use our first flowchart to tell us how to find the reduced row-echelon form of this
matrix:

Make
Place other
 leading 1 
Find R1 4 1 entries in 1
     
4 3 2 1 3 2 1 1 1 1 1 1 1
  and C1   in (1, 1)   C1 zero  
 1 1 1 1   1 1 1 1   4 3 2 1   0 −1 −2 −3 
Swap Add
       
0 1 2 3 0 1 2 3 R1 , R 2 0 1 2 3 −4R1 0 1 2 3
to R2
Find R2 , C2

Make Try to
Place other find more
 leading 1 
1 entries in 1
   
1 1 1 1 1 1 1 0 −1 −2 nonzero
  in (2, 2)   C2 zero   rows
 0 −1 −2 −3   0 1 2 3   0 1 2 3  Done!
Scale R2 Add −R2 None
     
0 1 2 3 by −1 0 1 2 3 to R1 , R3 0 0 0 0 exist!

We then consult our second flowchart to figure out what this means:

• We do not have any rows where all of the entries to the left of the vertical break are
0, but where the entry to the right is nonzero. So we are not in the “no solutions”
case.
• We now count our free variables. The variables x and y are leading, because there are
leading 1’s in their columns; this leaves z, which is our one free variable.
• According to our flowchart, this means we have a line of solutions!

Finally, if we want the equations for this line, we can just translate our augmented matrix
back into a set of equations:

 
1 0 −1 −2
x −z = −2
 0 1 2 3  −→
y +2z = 3
0 0 0 0

This gives us a set of general equations for our line. If we want a set of parametric equations,
we can set each free variable equal to its own parameter (in this case, set z = t) and rewrite
all of our other equations in terms of that parameter:

z = t,
x − t = −2 ⇒ x = t − 2,
y + 2t = 3 ⇒ y = 3 − 2t.

9 Matrix Operations and Applications [Lectures 16-19]

After reading and watching these lectures, we’re hoping that you can do the following tasks:

19
 • Add matrices together, multiply matrices by scalars, multiply matrices together, and
take the transpose of a matrix.
• Find the inverse of a matrix, and how to use it to find solutions to a system of linear
equations.
• Know what the identity matrix is.
• Know how to take the determinant of 2 × 2, 3 × 3 and n × n matrices in general.
• Know several useful properties of the determinant.

9.1 Matrix Arithmetic

 matrix of size m × n is a grid of numbers with m rows and n columns. For example,
A
1 2 3
is a 2 × 3 matrix. We defined several operations on matrices in class! Most of
2 3 1
these operations were very intuitive. Addition, for example, was pretty straightforward:
given two matrices A, B of the same size, we could create a new matrix A + B by just
adding each cell in A to the corresponding cell in B. For example,
       
1 2 3 0 1 0 1+0 2+1 3+0 1 3 3
+ = = .
2 3 1 1 0 1 2+1 3+0 1+1 3 3 2
 
1 2  
You can’t add matrices of different sizes; that is, + 1 2 4 DNE.
2 1
Scalar multiplication was similarly nice; given any m × n matrix A and any real number
x, we could form the matrix xA by multiplying each coordinate of A by x; for example,
   
1 3 3 2 6 6
2 = .
3 3 2 6 6 4

As well, given a matrix A we can form its transpose AT by “switching” its rows and
its columns; that is, given a matrix A, we can form the matrix AT by making a matrix
whose first row is A’s first column, whose second row is A’s second column, and so on/so
forth until we run out of columns. This is perhaps best illustrated by an example:
T 
1 2 3 1 4 7
 
4 5 6 = 2
  5 8
7 8 9 3 6 9

Not every matrix operation is intuitive, though. Matrix multiplication, as we saw in

class, is a pretty strange thing! We defined it as follows: suppose that A is a m × n matrix
and B is a n × l matrix. Then AB, the product of A and B, is defined, and in particular
is a m × l matrix! We defined this as follows: if we let rA,1 , . . . rA,m denote the m rows of
A and cB,1 , . . . cB,l denote the l columns of B, then
 
rA,1 · cB,1 rA,1 · cB,2 rA,1 · cB,3 . . . rA,1 · cB,l
 rA,2 · cB,1 rA,2 · cB,2 rA,2 · cB,3 . . . rA,2 · cB,l 
 
AB =  rA,3 · cB,1 rA,3 · cB,2 rA,3 · cB,3 . . . rA,3 · cB,l 
 
 .. .. .. .. .. 
 . . . . . 
rA,m · cB,1 rA,m · cB,2 rA,m · cB,3 . . . rA,m · cB,l

In other words, to find the entry that goes in (i, j), take the dot product of the i-th row of
A and the j-th column of B.
This probably looks scary, but in practice it’s not too bad. Here’s an example of this in
action:

20
    
1 2 1 −1 1
Example. If A =  2 3 −2 and D =  1 −1, find AD.
0 1 1 3 4

Answer. We use our definition as described above:

 

1 2 1

−1 1
 
rA,1 · cB,1 rA,1 · cB,2
 (1, 2, 1) · (−1, 1, 3) (1, 2, 1) · (1, −1, 4)
  
·c rA,2 · cB,2
   
 r
 =  A,2 B,1
−2  −1   =  (2, 3, −2) · (−1, 1, 3) (2, 3, −2) · (1, −1, 4) 
 2
AD =  3 1
  
   
0 1 1 3 4 rA,3 · cB,1 rA,3 · cB,2 (0, 1, 1) · (−1, 1, 3) (0, 1, 1) · (1, −1, 4)

   
−1 + 2 + 3 1−2+4 4 3
   
=
 −2 + 3 − 6  =  −5
2−3−8   −9 

0+1+3 0−1+4 4 3

Notice that to use our definition, we need the number of columns in A to be equal to
the number of rows in B. If A and B do not have the right  sizes to use
 the definition
 above,
1 2 3 1 2 3
we say that their product is undefined; so, for instance, DNE.
2 1 2 2 3 4
Matrix multiplication has a number of interesting properties. One is that most of
the time, the order of multiplication
  matters:
  that is, AB and BA are often very dif-
0 1 0 0 1 0
ferent! For example, = , but if we switch the order we can see that
0 0 1 0 0 0
    
0 0 0 1 0 0
= is quite different! Another property is that much like how R has
1 0 0 0 0 1
a “multiplicative identity” in 1 (that is, a number we can multiply by anything without
changing that thing), we have an identity for matrices as well: in general, for any n we
define the n × n identity matrix In as
 
1 0 0 ... 0
0 1 0 . . . 0
 
In = 0 0 1 . . . 0
 
 .. .. .. . . .. 
. . . . .
0 0 0 ... 1

This is a n × n matrix with ones on the main diagonal (i.e. 1’s in every cell (i, i))
and zeroes everywhere else. This matrix has the property
 that for any m × n matrix A,
1 2 3
Im · A = A · In = A. For example, if A = 4 5 6, then
7 8 9
 

1 0 0

1 2 3
 (1, 0, 0) · (1, 4, 7) (1, 0, 0) · (2, 5, 8) (1, 0, 0) · (3, 6, 9)
    
I3 A = 
 0 1  4
0  5 =
6  (0, 1, 0) · (1, 4, 7) (0, 1, 0) · (2, 5, 8) (0, 1, 0) · (3, 6, 9) 
  
 
0 0 1 7 8 9 (0, 0, 1) · (1, 4, 7) (0, 0, 1) · (2, 5, 8) (0, 0, 1) · (3, 6, 9)

 
1 2 3
 
=
 4 5 6 

7 8 9

9.2 Matrix Inverses

Given a matrix A, we call A square if it has as many rows as columns; in other words,
if it is size n × n for some n. Given any square matrix A, we say that A is invertible if

21
there is some matrix A−1 such that AA−1 = In = A−1 A; we call A−1 the inverse of A.
Not all matrices are invertible, but many are! In class, we described a process for finding
the inverse of a matrix A:

1. First, we construct the augmented matrix [A|I].

2. Then, we apply row operations to this augmented matrix to turn the left-hand side
into reduced row-echelon form.
3. If at the end of this process the left-hand side is the identity matrix I, then the
right-hand side is A−1 .
4. Otherwise, if the left-hand side is not the identity matrix after reducing it to its
reduced row-echelon form, then A−1 does not exist.

We look at an example here:

 
5 1 0
Example. Does the matrix C = 4 5 2 have an inverse?
5 3 1

Answer. We run our process here. First, we form [C|I]; then, we perform row operations
until the left-hand side is in RREF:
 
5 1 0 1 0 0
[C|I] =  4 5 2 0 1 0 
5 3 1 0 0 1
 
5−4 1−5 0−2 1−0 0−1 0−0
add −R to R , then add −R to R1
−−−−−−1−−−−3−−−−−−−−−−2−−−−→  4 5 2 0 1 0 
5−5 3−1 1−0 0−1 0−0 1−0
 
1 −4 −2 1 −1 0
= 4 5 2 0 1 0 
0 2 1 −1 0 1
 
1 −4 −2 1 −1 0
add −4R1 to R2
−−−−−−− −−−→  4 − 4 5 + 16 2+8 0−4 1+4 0+0 
0 2 1 −1 0 1
 
1 −4 −2 1 −1 0
=  0 21 10 −4 5 0 
0 2 1 −1 0 1
 
1 −4 −2 1 −1 0
add −10R3 to R2
−−−−−−−− −−−→  0 − 0 21 − 20 10 − 10 −4 + 10 5 0 − 10 
0 2 1 −1 0 1
 
1 −4 −2 1 −1 0
= 0 1 0 6 5 −10 
0 2 1 −1 0 1
 
1+0 −4 + 4 −2 + 0 1 + 24 −1 + 20 0 − 40
add −2R to R ,4R →R1
−−−−−−−2−−−−3−−−2−−−→  0 1 0 6 5 −10 
0−0 2−2 1−0 −1 − 12 0 − 10 1 + 20
 
1 0 −2 25 19 −40
= 0 1 0 6 5 −10 
0 0 1 −13 −10 21
 
1 0 −2 + 2 25 − 26 19 − 20 −40 + 42
add 2R3 to R1
−−−−−− −−−→  0 1 0 6 5 −10 
0 0 1 −13 −10 21
 
1 0 0 −1 −1 −2
= 0 1 0 6 5 −10 
0 0 1 −13 −10 21

22
 The left-hand side
 is in reduced row-echelon
 form, and in fact is the identity; therefore
−1 −1 2
the right-hand side  6 5 −10 is C −1 !
−13 −10 21
To make sure we haven’t made any errors in our calculations, we check that CC −1 is in
fact equal to I here:

 

5 1 0

−1 −1 2
 (5, 1, 0) · (−1, 6, −13) (5, 1, 0) · (−1, 5, −10) (5, 1, 0) · (2, −10, 21)
    
CC −1 = 
 4 5  6
2  5 =
−10  (4, 5, 2) · (−1, 6, −13) (4, 5, 2) · (−1, 5, −10) (4, 5, 2) · (2, −10, 21) 
  
 
5 3 1 −13 −10 21 (5, 3, 1) · (−1, 6, −13) (5, 3, 1) · (−1, 5, −10) (5, 3, 1) · (2, −10, 21)

   
−5 + 6 + 0 −5 + 5 + 0 10 − 10 + 0 1 0 0
   
=
 −4 + 30 − 26 −4 + 25 − 20 8 − 50 + 42 = 0
 1 0 

−5 + 18 − 13 −5 + 15 − 10 10 − 30 + 21 0 0 1

Success!
If this matrix was not invertible,
  we would not have gotten the identity on the left at
2 1
the end. For example, B = is not invertible, because
8 4
1 1
   
2 1 1 0 multiply R1 by 1
2 1 0
[B|I] = −−−−−−−−−−−→ 2 2
8 4 0 1 8 4 0 1
1 1 1 1
   
add−8R1 to R2 1 0 1 0
−−−−−−−−−−→ 2
1
2
1 = 2 2
8−8·1 4−8· 2
0−8· 2
1−8·0 0 0 −4 1

gives us a matrix whose left-hand side is in RREF but is not the identity I2 .
Inverses of matrices can be used to solve systems of linear equations! Notice that if we
have n linear equations in n unknowns, like for example
5x+ y = 1,
4x+ 5y +2z = 1,
5x+ 3y +z = 1.
we can rewrite this as
    
5 1 0 x 1
4 5 2 y  = 1 .
5 3 1 z 1
In general, if you have a system of linear equations in n variables, if you let A be the
matrix of coefficients of those variables, x be the vector consisting of all of those variables,
and b be the vector of the constants each equation is equal to, you can always express that
system of linear equations as Ax = b, just like we’ve done here.  
5 1 0
Returning to this example: earlier in these notes, we said that if C = 4 5 2 then
5 3 1
 
−1 −1 2
C is invertible, and in particular C −1 =  6 5 −10 .
−13 −10 21
   
x a
Therefore, if we want to solve the equation C y  =  b , we can just multiply both
z c
sides by C −1 to get

23
  
   

−1 −1 2

1
 (−1, −1, 2) · (1, 1, 1) 
−1 − 1 + 2
 
0

x a     
−1
   
 1  = 
=
= 1 
y  = C  b  =  6 5 −10    (6, 5, −10) · (1, 1, 1)   6 + 5 − 10   

z c  
−13 −10 21 1 (−13, −10, 21) · (1, 1, 1) −13 − 10 + 21 −2

   
x 1
In other words, we’ve solved our system of linear equations C y = 1, and found that
  
z 1
     
x 1 0
y  = C −1 1 =  1 !
z 1 −2
This process works in general: if you have a system of linear equations of the form
Ax = b, then if A−1 exists, you get exactly one solution to this system, and it’s A−1 b! In
general, this is not the fastest way to solve a system of linear equations, and it only applies
when you have A−1 ; if A−1 does not exist, then you cannot use this method, and should go
back to our earlier methods using the RREF to find a solution. But if someone has given
you A−1 for free, then this is a faster way to solve systems of linear equations!
In particular, this means that the inverse is connected to finding solutions to systems of
linear equations in certain ways:

• Let A be a square matrix. If A−1 exists, then Ax = b has exactly one solution for
every b.
• This also applies in the other direction: if A is a square matrix and Ax = b has
exactly one solution for some b, then A−1 exists.

9.3 The Determinant

Finally, the last operation we described for matrices was the determinant.
 The
 determi-
1 2 4
nant det(A) is something only defined for square matrices (so det does not
2 5 0
exist), and we defined it as follows:

• For a 1 × 1 matrix [a], det([a])

 = a.  
a b a b
• For a 2 × 2 matrix , we have det = ad − bc.
c d c d
   
a b c a b c
• For a 3 × 3 matrix d e f , we have det d e f  = aei + bf g + cdh − af h −
g h i g h i
bdi − ceg. A nice way to remember this 3 × 3 determinant formula is to write the
matrix next to itself, and then circle the six diagonal lines labeled below:

  
a b c a b c
d e f d e f 
g h i g h i

The three blue diagonals correspond to the three terms you add, and the three red
diagonals are the three terms you subtract in the formula above.

For larger matrices, like 4 × 4 and on up, most of the formulas you could memorize get
very messy very quickly. So instead we came up with some properties that can help you
calculate the determinant of a large matrix quickly:

• Given a square matrix A, we know how our row operations from earlier affect the
determinant of A:

24
 – If we multiply a row of A by a constant c, this multiplies the determinant by c.
– If we switch two rows in A, this multiplies the determinant by −1.
– If we add a multiple of one row to another row in A, this does nothing to the
determinant.
• We say that a matrix A is upper-triangular if the only cells in A that contain
nonzero values are those on or above the
 main diagonal; that is, upper-triangular
1 0 2
matrices are ones that look like 0 2 3.
0 0 0
• The determinant of a matrix that is upper-triangular is the product of the entries on
its diagonal.
Accordingly, this gives us a nice blueprint for how to find the determinant of any square
matrix A:
• Take A and perform row operations on it to transform it into an upper-triangular
matrix B.
• Calculate the determinant of B by multiplying the entries on its diagonal!
• Use this to find the determinant of A by correcting for the row operations you per-
formed: that is, for each swap you did to A, make sure to multiply det(B) by −1 to
cancel out the earlier −1, and for each time you multiplied a row in A by a constant
c make sure to multiply det(B) by 1c .
We calculate a few examples here:
Example. Find the determinants of the following matrices:
 
  1 2 0 0 0
  9 8 7 2 1 2 0 0
7 2  
A= , B = 6 5 4 , D = 0 2 1 2 0
2 1 
3 2 1 0 0 2 1 2
0 0 0 2 1
Answer. For A and B, we just use the formula for the determinant of a 2 × 2 matrix:
 
7 2
det =7 · 1 − 2 · 2 = 3,
2 1
    
9 8 7 9 8 7 9 8 7
det 6 5 4 = 6 5 46 5 4
3 2 1 3 2 1 3 2 1
=9 · 5 · 1 + 8 · 4 · 3 + 7 · 6 · 2 − 9 · 4 · 2 − 8 · 6 · 1 − 7 · 5 · 3
=45 + 96 + 84 − 72 − 48 − 105 = 0.

For D, we use row operations to transform this matrix into a triangular matrix:
     
1 2 0 0 0 1 2 0 0 0 1 2 0 0 0
2
 1 2 0 0 add −2R1 to R2 
0 −3 2 0 0 2
 add 3 R1 to R2 
0 −3 2 0 0

7
 −−−−−−−−−−→ 0  −−−−−−−−−−→ 0
0 2 1 2 0  2 1 2 0  0 2 0
 3 
0 0 2 1 2 0 0 2 1 2 0 0 2 1 2
0 0 0 2 1 0 0 0 2 1 0 0 0 2 1
   
1 2 0 0 0 1 2 0 0 0
6
add − R1 to R2
0
 −3 2 0 0
 add 14 R1 to R2
0 −3
 2 0 0
7 7
−−−−−7−−−−−−→  0 0 3
2 0 5
 −−−−−−−−−−→ 0
 0 3
2 0
5
0 0 0 − 7 2 0 0 0 − 57 2
33
0 0 0 2 1 0 0 0 0 5

25
 The determinant of the matrix at right is just the product of the entries on its diagonal,
i.e. 1 · (−3) · 73 · (− 75 ) · 33
5 = 33, because it is upper-triangular. Therefore, because adding
rows to other rows does not change the determinant, the determinant of the original matrix
is also 33.

The determinant has some nice properties:

• A square matrix A has determinant 0 if and only if A is not invertible.

• If a square matrix A has two identical rows or columns, or a row of all zeroes, then
det(A) = 0.
• If A, B are both square n × n matrices, then det(AB) = det(A) det(B).
• det(AT ) = det(A).

10 Cross Product [Lecture 20]

After reading and watching these lectures, we’re hoping that you can do the following tasks:

• Take the cross product of two vectors in R3 .

• Know several properties about the cross product.

10.1 Definition and Properties

The cross product is an operation that takes in two vectors in R3 and outputs another
vector in R3 . It is defined as follows: for any (u1 , u2 , u3 ), (v1 , v2 , v3 ) ∈ R3 , we have

(u1 , u2 , u3 ) × (v1 , v2 , v3 ) = (u2 v3 − u3 v2 , u3 v1 − u1 v3 , u1 v2 − u2 v1 ).

For example,

(1, 2, 3) × (0, 4, 5) = (2 · 5 − 3 · 4, 3 · 0 − 1 · 5, 1 · 4 − 2 · 0) = (−2, −5, 4).

The cross product has several useful properties:

• Given any two vectors u, v, u × v is orthogonal to both u and v.

• As well, ||u × v|| is equal to the area of the parallelogram spanned by u and v.
• Finally, ||u × v|| = ||u||||v|| sin(θ), where θ is the angle between u and v.

11 Differentiation [Lectures 21-23]

After reading and watching these lectures, we’re hoping that you can do the following tasks:

• Know the definition of the derivative and how to use it.

• Find the derivatives of various basic functions.
• Take the derivatives of more complex functions via the product and chain rules.
dy
• Use implicit differentiation to find dx of a relation involving x and y.
• Find the tangent line to a curve.

11.1 Calculating Derivatives

dy
The derivative of a function f at some point x, denoted as either f 0 (x) or d
dx f (x) or dx if
we’re looking at the graph y = f (x), is the following limit:

f (x + h) − f (x)
f 0 (x) = lim
h→0 h

26
So, for example, the derivative of f (x) = x2 at x = 2 is just

(2 + h)2 − 22 4 + 4h + h2 − 4 4h + h2
lim = lim = lim = lim 4 + h = 4.
h→0 h h→0 h h→0 h h→0

Geometrically, we think of the derivative of a function f at some point x as measuring the

“slope” of our function at that point. We can visualize this by drawing the tangent line
to a function f (x) at some point a, which has the equation

y − f (a) = f 0 (a) · (x − a).

So, if we return to our example above where f (x) = x2 , we can see that a tangent line to
f (x) at x = 2 would have the equation

y − f (2) = f 0 (2)(x − 2) ⇒ y − 4 = 4(x − 2).

Drawing this line next to the graph of y = f (x) shows that we are indeed capturing the
idea of the “slope” of our function at x = 2:

10

-2 -1 0 1 2 3 4 5 6

While elegant, this limit definition of the derivative can take a while to use. Accordingly,
we’ve calculated the derivatives of several simple functions:
d x x d n n−1 , d
• dx e = e • dx x = nx n 6= 0 • dx cos(x) = − sin(x)
d 1 d d
• dx ln(x) = x • dx c = 0 • dx sin(x) = cos(x)

We also have a set of rules that let us take the derivative of more complicated functions:

• Differentiation is linear; given any two functions f (x), g(x) and constants a, b, we
d
have dx (af (x) + bg(x)) = af 0 (x) + bg 0 (x).
• Product rule: given any two functions f (x), g(x), we have dx d
(f (x) · g(x)) = f 0 (x) ·
0
g(x) + f (x) · g (x).
d
• Chain rule: given any two functions f (x), g(x), we have dx (f (g(x))) = f 0 (g(x))·g 0 (x).

We look at a few quick examples of these rules in action:

√
Example. Find the derivatives of p(x) = e x , q(x) = sin(x) cos(x) and r(x) = x2 ·ln(x2 +1).

Answer. For p(x), we want to use the chain rule; this is because p(x) consists of functions
composed with each other, and the chain rule is the only rule that deals with this! So: let’s
√
write p(x) = f (g(x)), where f (x) = ex and g(x) = x. Then, the chain rule tells us that
p0 (x) = dxd
(f (g(x))) = f 0 (g(x)) · g 0 (x). We know from above that f 0 (x) = dx
d x
e = ex , and
√
that g 0 (x) = dxd
x = 21 x−1/2 = 2√1 x ; therefore, we have
d 1/2
x = dx
√
0 0 0
√
x 1 e x
p (x) = f (g(x)) · g (x) = e · √ = √ .
2 x 2 x

For q(x), we want to use the product rule, because q(x) consists of the product of two
functions. If we let f (x) = sin(x), g(x) = cos(x) then q(x) = f (x)g(x); so the product rule

27
 d
says that dx q(x) = dx d
(f (x) · g(x)) = f 0 (x)g(x) + f (x)g 0 (x). Because f 0 (x) = d
dx sin(x) =
cos(x) and g 0 (x) = dx
d
cos(x) = − sin(x), this tells us that

q 0 (x) = f 0 (x)g(x) + f (x)g 0 (x) = cos2 (x) − sin2 (x).

r(x) might seem harder to determine which rule to use, but it’s actually not that bad: if
we look at r(x), we just need to decide whether it looks more like a f (g(x)) or a f (x)g(x)!
In this case, it’s not clear how we would write this as a f (g(x)), as there’s not an obvious
“outside” function that we’re applying to some inside function. However, it’s very easy to
see how we’d write this as a product: we can write r(x) = f (x) · g(x), where f (x) = x2 and
g(x) = ln(x2 + 1). This is how differentiation always works; you’ll always have exactly one
rule that can work, and all you have to do is figure out what that rule is and then apply it!
If we do that here, then f 0 (x) = dx
d 2
x = 2x, while g 0 (x) = dx d
ln(x2 + 1) is trickier; here
we have to use the chain rule, because we have one function (ln(x)) being applied to another
(x2 + 1)! In particular, if we let h(x) = ln(x), j(x) = x2 + 1, then h0 (x) = x1 , j 0 (x) = 2x, and
therefore the chain rule tells us that g 0 (x) = h0 (j(x)) · j 0 (x) = x22x+1 .
Plugging this into our product rule work earlier tells us that

2x3
r0 (x) = f 0 (x)g(x) + f (x)g 0 (x) = 2x ln(x2 + 1) + .
x2 + 1

11.2 Implicit Differentiation

Sometimes we will want to find the tangent line to an equation like

2x(x2 + y 2 ) = 3x2 − y 2

even though we cannot easily solve for y and make this into a function of x! In this situation,
dy
we use implicit differentiation to try to find dx . The idea here is the following: take any
expression involving the variables x and y, like sin(x) or exy or y 2 − 2x + 1 or ln(y).
d
• If this expression has the form f (x) (in other words, it only involves x), define dx f (x) =
0
f (x). In other words, take the derivative like normal.
d
• If this expression has the form f (y) (in other words, it only involves y), define dx f (y) =
0 dy dy
f (y) · dx . In other words, take the derivative like normal, but stick this dx on the
outside.
• If it has both x’s and y’s, use the chain and product rules to break it into smaller
pieces.

Given an equation in two variables x, y, implicit differentiation is the following pro-

cess:
d
• Apply dx to both sides of this equation, as described above.
dy dy
• This gives you an equation with x’s, y’s, and dx ’s. Solve for dx in terms of the other
dy
variables by putting the dx ’s on one side and all of the other terms on the other side.
dy
• You now have an expression for dx in terms of x and y: in other words, you have
implicitly differentiated your equation!

To give an example: let’s look at the curve 2x(x2 + y 2 ) = 3x2 − y 2 from earlier.

28
 d
If we apply dx to both sides, we get

d d
2x · (x2 + y 2 ) = 3x2 − y 2



  dx   dx
d d d d 2
⇒ (2x) · (x2 + y 2 ) + (2x) · (x2 + y 2 ) = (3x2 ) − (y )
dx dx dx dx
 
d 2 d 2 dy
⇒ 2(x2 + y 2 ) + (2x) (x ) + (y ) = 6x − 2y
dx dx dx
 
2 2 dy dy
⇒ 2(x + y ) + (2x) 2x + 2y = 6x − 2y
dx dx
dy dy
⇒ 6x2 + 2y 2 + 4xy = 6x − 2y .
dx dx

dy
Now, we solve for :
dx
dy dy
6x2 + 2y 2 + 4xy = 6x − 2y
dx dx
dy dy
⇒ 4xy + 2y = 6x − 6x2 − 2y 2
dx dx
dy
⇒ (4xy + 2y) = 6x − 6x2 − 2y 2
dx
dy 6x − 6x2 − 2y 2
⇒ =
dx 4xy + 2y

To check
 that our answer makes sense, let’s try graphing a tangent line to this curve at
dy
the point 1, √13 . Plugging this point into our equation for dx yields
 2
6 · 1 − 6 · 12 − 2 · √1
dy 3 − 23 1
= = =− √ ,
dx 4·1· √1 + 2 √13 √6 3 3
3 3

 
which tells us that a tangent line to our curve at 1, √13 has equation

dy 1 1
y − y0 = (x − x0 ) ⇒ y − √ = − √ (x − 1).
dx 3 3 3
Graphing this line verifies that it is indeed a tangent line:

29
12 Differentiation Applications [Lectures 24-26]
After reading and watching these lectures, we’re hoping that you can do the following tasks:

• Know what it means for a function to be increasing, decreasing, concave up, concave
down, to have an inflection point, a critical point, or a relative maxima or minima.
• Visually identify all of the above properties.
• Use the derivative to find where a function has any of the above properties.

12.1 Definitions
The derivative can help us visualize and draw functions! It does this in many ways:

• Given a function f , we say that f is increasing on the interval (a, b) if for any
x < y ∈ (a, b), we have f (x) < f (y). Similarly, we say that f is decreasing on (a, b)
if for any x < y ∈ (a, b) we have f (x) > f (y).

Increasing Decreasing

• The derivative can tell us when this happens! It turns out that f is increasing on
(a, b) if f 0 (x) > 0 for every x ∈ (a, b), and f is decreasing on (a, b) if f 0 (x) < 0 for
every x ∈ (a, b).
• We say that f is concave up on (a, b) if f 00 (x) > 0 on (a, b); similarly, f is concave
down on (a, b) if f 00 (x) < 0 on (a, b). Visually, concave up graphs look like they’re
curving upwards (think cups, rockets taking off, the parabola y = x2 ) and concave
down graphs look like they’re curving downwards (think waterfalls, the path made by
throwing a ball in the air, y = −x2 .)

Concave up Concave down

• We say that x is a critical point if f 0 (x) = 0 or does not exist.

• We say that a point x ∈ (a, b) is a maximum on (a, b) if f (x) ≥ f (y), for any
y ∈ (a, b). We say that x is a relative maximum if there is some interval containing
x such that x is a maximum on that interval. Similarly x is a minimum on (a, b) if
f (x) ≤ f (y) for any y ∈ (a, b), and is a relative minimum if there is some interval
containing x in which x is a minimum.

30
 • The derivative can help us find these objects! If x is a relative maxima or minima,
then f 0 (x) is a critical point. Not all critical points are relative maxima or minima,
but all relative maxima and minima are critical points.
• We say that a point a is a point of inflection if f 00 (x) changes from positive to negative,
or vice-versa, at x = a.

Inflection point

We can use these properties to draw remarkably accurate graphs of functions! We look
at an example here:
Example. Draw the graph of f (x) = x(x − 9)(x − 24) , labeling all critical points, relative
maxima and minima, inflection points, and identifying where the function is concave up
and where it is concave down.
Answer. Our process for drawing a graph is as follows:
• We start by finding all of the places where our function crosses the x-axis and y-axis:
in other words, we find all of the values of x for which f (x) = 0, and also what f (0)
is.
• Then, we find f 0 (x), and find out where it is positive and negative. We use this to
identify all of the critical points of f and identify which are minima and which are
maxima; we also use this to determine where f is increasing and where f is decreasing.
• We finish by finding f 00 (x), and determine where this is positive and where this is
negative; we use this to find the inflection points of f , and determine where f is
concave up and concave down.
The first of these tasks is pretty straightforward. We know that f (x) = x(x − 9)(x − 24);
so we’ve already factored x into its roots, and can see that we have f (x) = 0 whenever x
is 0, 9 or 24. Similarly we know that f (0) = 0(0 − 9)(0 − 24) = 0, so we know where our
function crosses the y-axis.
Now, to get some more information we look at f 0 (x). Because
f (x) = x(x − 9)(x − 24) = x3 − 33x2 + 216x
⇒ f 0 (x) = 3x2 − 66x + 216 = 3(x2 − 22x + 72) = 3(x − 4)(x − 18),
we can see that x = 4, 18 are the two critical points of our function f . Moreover, because
3(x − 4) (x − 18) 3(x − 4)(x − 18)
x ∈ (−∞, 4) (−) (−) (−) · (−) = (+)
x ∈ (4, 18) (+) (−) (+) · (−) = (−)
x ∈ (18, ∞) (+) (+) (+) · (+) = (+)
we can see that our function is increasing on (−∞, 4), decreasing on (4, 18) and then in-
creasing again on (18, ∞). Finally, because at 4 we switch from increasing to decreasing we
know that our function has a relative maximum there, and at 18 because we switch from
decreasing to increasing we have a relative minimum.
This gives us some more information about our function! In particular, we can plot the
points
f (4) = 4(4 − 9)(4 − 24) = 400, f (18) = 18(18 − 9)(18 − 24) = −972
and get that our function looks like something that goes through the following points,
increasing until x = 4, decreasing until x = 18, and then increasing again:

31
 400

200

-5 0 5 10 15 20 25 30

-200

-400

-600

-800

-1000

Finally we look at f 00 (x):

d 0 d
f 00 (x) = 3x2 − 66x + 216 = 6x − 66 = 6(x − 11).


f (x) =
dx dx
This is negative for all x < 11 and positive for all x > 11; therefore our function is concave
down on (−∞, 11) and concave up on (11, ∞), with an inflection point at x = 11.
At x = 11 we have f (11) = −286, so this gives us one last point to plot.
Now, we draw! Specifically, we draw a nice concave-down curve through the points
(0, 0), (4, 400), (11, −286) and a nice concave-up curve through the points (11, −286), (18, −972), (24, 0):
400

200

-5 0 5 10 15 20 25 30

-200

-400

-600

-800

-1000

13 Integration [Lectures 27-30]

After reading and watching these lectures, we’re hoping that you can do the following tasks:
• Know the definition of the antiderivative, and the antiderivative of several basic func-
tions.
• Use integration by substitution and integration by parts to find more complicated
integrals.
• Know how to take a definite integral by using the fundamental theorem of calculus.
• Find the area between a curve y = f (x) and the x-axis.

13.1 The Indefinite Integral

Given a function f , we say that F is an antiderivative of f if F 0 (x) = f (x). In this sense,
taking an antiderivative is exactly what it sounds like: it’s just “undoing” the derivative!
A given function may have many antiderivatives. For instance, f (x) = cos(x) has
d
F (x) = sin(x) + 9 as an antiderivative, because dx (sin(x) + 9) = cos(x); but it also has
sin(x) − 42, sin(x), and sin(x) + π as antiderivatives as well, because the constant term
d
doesn’t matter if we’re applying dx !
Accordingly, we define the indefinite integral of a function f as “the antiderivative
up to a constant:” that is, if f (x) is a function and F (x) is any antiderivative of f (x), we
write
Z
f (x) dx = F (x) + C, C∈R

to denote the indefinite integral of f (x).

We know the indefinite integrals of several basic functions:

32
 Z
xn+1
Z
• ex dx = ex + C • xn dx = + C, n 6= −1
Z Z n+1
• ln(x) dx = x ln(x) − x + C • cos(x) dx = sin(x) + C
Z
1
Z
• dx = ln(x) + C • sin(x) dx = − cos(x) + C
x

We also have some techniques for integrating more complicated functions! One technique
is integration by substitution, which you can think of as “reverse chain rule.” It’s the
following formula:
Z Z
0 0 d
f (g(x)) · g (x) dx = (f (g(x))) dx = f (g(x)) + C, C ∈ R.
dx
Basically, this technique says that if we can recognize the function we’re integrating as
having the form f 0 (g(x)) · g 0 (x) for some f, g, then we’re automatically done! We just get
that the integral is f (g(x)) + C, and that’s quite nice.
Sometimes people use “u-substitution” notation, where to evaluate the integral
Z
f 0 (g(x)) · g 0 (x) dx

they define u = g(x), which means that g 0 (x) = du 0

dx , and therefore that du = g (x)dx.
Substituting these u’s in for x’s gives us
Z Z
f (g(x)) · g (x) dx = f 0 (u) du = f (u) + C = f (g(x)) + C;
0 0

in other words it does the exact same thing as the notation above. Pick your favorite!
Not all functions can be written in the form f 0 (g(x)) · g 0 (x), though. For those other
kinds of functions, we have integration by parts (which we can think of as “reverse product
rule.”) It’s the following formula:
Z Z
f (x) · g(x) dx = f (x)g(x) − f (x)g 0 (x) dx.
0

Basically, this technique says that if we can write the function

R 0 we’re integrating as a product
f 0 (x)g(x), then we can “trade” the problem of finding f (x)g(x) dx for the problem of
0
R
finding f (x)g (x) dx. This can make our problems a lot easier; if we make g(x) something
like ln(x) or x, for instance, g 0 (x) gets a lot nicer when we take a derivative! Conversely, if
we make f 0 (x) something easy to integrate, like ex or sin(x), then f (x) won’t be any worse,
and in theory we will have exchanged a tricky integral for an easier one.
This makes a lot more sense in practice, so let’s look at some examples:

Example. Using the techniques of integration by parts and integration by substitution (i.e.
the reverse chain and product rules), find each of the following indefinite integrals:
Z
1. (x + 1) sin(x + 1) dx
Z
2. (x + 1) sin((x + 1)2 ) dx
Z
sin(x)
3. dx
Z cos(x)
4. (ln(x))2 dx

33
 Z
Answer. 1. The first thing we need to do to evaluate (x+1) sin(x+1) dx is figure out
which technique we want to try. At first glance, reverse chain rule (i.e. integration by
substitution) looks good, in that we have some composition going on here — we’d be
tempted to make f 0 (x) = sin(x) and g(x) = x + 1. However, the thing on the outside
is 0
Z not g (x) = 1; it’s x + 1! So our integral actually doesn’t look like it’s of the form
g 0 (x)f 0 (g(x)) dx, and as a result we’re better off trying something else.
Z
Let’s try parts, then! If we were to apply integration by parts to (x+1) sin(x+1) dx,
we’d want to write this in the form f 0 (x)g(x)dx, where f 0 (x) is something whose
R

integral we know and isn’t too bad, while g(x) is something that hopefully gets simpler
when we differentiate. This motivates us to choose g(x) = x + 1, because g 0 (x) = 1 is
indeed a lot simpler; this leaves us with f 0 (x) = sin(x + 1), which has the reasonable
integral f (x) = − cos(x + 1).
Integration by parts, then, tells us that
Z Z Z
(x + 1) sin(x + 1) dx = f (x) · g(x) dx = f (x)g(x) − f (x)g 0 (x) dx
0

Z
= (− cos(x + 1))(x + 1) − (− cos(x + 1))(1) dx
Z
= −(x + 1) cos(x + 1) + cos(x + 1) dx

= −(x + 1) cos(x + 1) + sin(x + 1) + C, C ∈ R.

We can check that −(x + 1) cos(x + 1) + sin(x + 1) + C is indeed an antiderivative of

(x + 1) sin(x + 1) by using the product rule:

d
(−(x + 1) cos(x + 1) + sin(x + 1) + C)
dx
= − cos(x + 1) + (−(x + 1))(− sin(x + 1)) + cos(x + 1)
=(x + 1) sin(x + 1).
Z
2. The integral (x + 1) sin((x + 1)2 ) dx looks like a much better integration by sub-
stitution candidate! As before, we think of sin((x + 1)2 ) as the “f 0 (g(x))” part,
with f 0 (x) = sin(x) and g(x) = (x + 1)2 ; this now means that f (x) = − cos(x),
g 0 (x) = 2(x + 1), and therefore that
Z Z
2 1
(x + 1) sin((x + 1) ) dx = 2(x + 1) sin((x + 1)2 ) dx
2
Z
1
= f 0 (g(x)) · g 0 (x) dx
2
1
= f (g(x)) + C
2
1
= − cos((x + 1)2 ) + C, C ∈ R.
2

34
 As always, we check this antiderivative by taking a derivative, using the chain rule:
 
d 1 1 d
− cos((x + 1) ) + C = − (− sin((x + 1)2 )) ·
2
(x + 1)2
dx 2 2 dx
1
= − (− sin((x + 1)2 ))(2(x + 1)1 )
2
= (x + 1) sin((x + 1)2 ).

Z
sin(x)
3. If we look at dx, it looks like substitution is not a bad guess: we certainly
cos(x)
have some composition going on with the cos(x)1
part, and if we indeed make f 0 (x) =
1 0
x , g(x) = cos(x) then g (x) = − sin(x) does indeed give us the remaining parts of
the function we’re integrating, up to the sign! Therefore, because f 0 (x) = x1 forces
f (x) = ln(x), we have
Z Z
sin(x) 1
dx = − − sin(x) dx
cos(x) cos(x)
Z
= − f 0 (g(x)) · g 0 (x) dx

= −f (g(x)) + C
= − ln(| cos(x)|) + C, C ∈ R.

sin(x)
We can check that this is indeed the antiderivative of tan(x) = cos(x) by using the
chain rule:
d 1 d
(− ln(| cos(x)|) + C) = − · (| cos(x)|)
dx | cos(x)| dx
(
1 d
− cos(x) · dx (cos(x)), if cos(x) ≥ 0
= 1 d
− − cos(x) · dx (− cos(x)), if cos(x) < 0
(
1
− cos(x) · (− sin(x)), if cos(x) ≥ 0
= 1
− − cos(x) · (−(− sin(x))), if cos(x) < 0
( sin(x)
cos(x) , if cos(x) ≥ 0
= sin(x)
cos(x) , if cos(x) < 0
sin(x)
= .
cos(x)
Z
4. If we were to try integration by substitution on (ln(x))2 dx , we’d have to make
f 0 (g(x)) = (ln(x))2 , and this doesn’t really leave anything left for the g 0 (x) part! So,
let’s not try
Z this, and try parts instead! In particular, this means we probably think
of this as ln(x) ln(x) dx .
This makes our choice for f 0 (x) and g(x) pretty simple: we make f 0 (x) = ln(x)
and g(x) = ln(x), because we don’t have any other choices really! This means that

35
 f (x) = x ln(x) − x, as we saw in class earlier, while g 0 (x) = x1 . As a result, we have
Z Z Z
ln(x) ln(x) dx = f (x) · g(x) dx = f (x)g(x) − f (x)g 0 (x) dx
0

Z
1
= (x ln(x) − x) ln(x) − (x ln(x) − x) dx
x
Z
= x(ln(x))2 − x ln(x) − ln(x) − 1 dx

= x(ln(x))2 − x ln(x) − (x ln(x) − x − x) + C

= x(ln(x))2 − 2x ln(x) + 2x + C.

We can check that x(ln(x))2 − 2x ln(x) + 2x + C is indeed an antiderivative of (ln(x))2

by using the product and chain rules:
d
x(ln(x))2 − 2x ln(x) − 2x + C


dx
1
=(ln(x))2 + x · 2 ln(x) · − 2 ln(x) − 2 + 2
x
2
=(ln(x)) ,

13.2 Definite Integration

The definite integral of a function f (x) from a to b, written
Z b
f (x) dx,
a

is the signed area between the curve y = f (x) and the x-axis, where we think of area
above the x-axis as being positive and area below the x-axis as being negative. So, for
instance, the definite integral
Z 2π
sin(x) dx = 0,
0

because the area above the curve from 0 to π is “canceled out” by the area below the curve
from π to 2π:

+
-

In general, though, we calculate definite integrals by using the Fundamental Theorem

of Calculus: if f (x) is a function and F (x) is any antiderivative of f (x), then
Z b
f (x) dx = F (b) − F (a).
a

So, for example, because − cos(x) is an antiderivative of sin(x), we have

Z 2π
sin(x) dx = (− cos(2π)) − (− cos(0)) = −1 − (−1) = 0,
0

36
which verifies algebraically the fact we geometrically saw a moment ago.
One useful application of the definite integral is to finding the unsigned area between
a curve and the x-axis: i.e. the area where we count area below and above the x-axis as
positive, and don’t have any of this canceling-out stuff! To find this, you just want to
integrate |f (x)|, as the absolute-value signs transform all of the parts of our curve where
they were below the x-axis into parts that are above the x-axis! In other words, if we let
A denote the unsigned area between the x-axis and the curve y = f (x) from a to b, then
Rb
A = a |f (x)| dx.
To integrate something like |f (x)|, it helps to break up the region you’re integrating
f (x) over into places where f (x) ≥ 0 and where f (x) ≤ 0, so that you can replace
We illustrate this idea with an example:
Z 1
x
Example. Find f (x) dx for f (x) = 2 . Then, find the area between the x-axis
−1 x +3
and the curve y = f (x) from x = −1 to x = 1.
Z
x
Answer. We start by finding the indefinite integral dx. This looks like an in-
x2 + 3
tegration by substitution problem, as it has some composition going on; indeed, if we let
f 0 (x) = x1 , g(x) = x2 + 3, g 0 (x) = 2x we have f (x) = ln(|x|) and therefore that
Z Z Z
x 1 x 1
dx = 2 2 dx = g 0 (x)f 0 (g(x)) dx
x2 + 3 2 x +3 2
1
= f (g(x)) + C
2
1
= ln(|x2 + 3|) + C.
2
Therefore, by the fundamental theorem of calculus, we have
Z 1    
x 1 1
dx = ln(|x2 + 3|) + C − ln(|x2 + 3|) + C
−1 x2 + 3 2 2
x=1 x=−1
ln(4) + C ln(4) + C
= −
2 2
= 0.

This answers the first part of our problem. Z

x
To answer the second, we need to figure out dx. These absolute value signs
+3 x2
are pretty irritating, so we want to get rid of them! To do this, we need to figure out where
our curve is above the x-axis and where it is below the x-axis.
x
Notice that for any x, x2 + 3 > 0; so the only part of 2 relevant to determining the
x +3
sign of our function is the numerator, which is x. Therefore this entire function is positive
when x > 0 and negative when x < 0!
x x x x
As a result, we can see that 2 = 2 when x ≥ 0, and 2 =− 2
x +3 x +3 x +3 x +3
when x < 0; as a result, we can write

37
 Z 1 Z 1 Z 0
x x x
Area = 2
dx = 2
dx + 2
dx
−1 x +3 0 x +3 −1 x + 3
Z 1 Z 0
x x
= 2
dx + − 2 dx
0 x +3 −1 x +3
    !
1 2 1 2
= ln(|x + 3|) + C − ln(|x + 3|) + C
2 2
x=1 x=0
    !
1 1
− ln(|x2 + 3|) + C − ln(|x2 + 3|) + C
2 2
  x=0  x=−1
ln(4) + C ln(3) + C ln(3) + C ln(4) + C
= − − −
2 2 2 2
= ln(4) − ln(3).

14 Functions of Two Variables [Lectures 31-32]

After reading and watching these lectures, we’re hoping that you can do the following tasks:
• Find the level curves of a function f (x, y), and use them to visualize the graph z =
f (x, y).
d d
• Calculate partial derivatives dx f (x, y), dy f (x, y).
• Find the tangent plane to a function f (x, y) at a point.

14.1 Functions of Two Variables

A function of two variables f : D → R, for any subset D of R2 , is any rule that takes in
pairs (x, y) ∈ R2 of real numbers and outputs another real number. We think of D above
y
as the domain of this function. As before, given a rule f like f (x, y) = x+1 , we often want
to find sets on which our rule is a function (and in particular is defined); for this example,
for instance, f is defined on the set {(x, y) | x, y ∈ R, x 6= −1}.
Pretty much all of the functions of this sort that we interact with in this course are those
that we make by sticking together various elementary functions that we know: for example,
f (x, y) = sin(xy), g(x, y) = x2 + y 2 , h(x, y) = ex − y are all functions of two variables.
Given a function f of two variables, we often want to graph that function! These graphs
take place in three dimensions, i.e. R3 , and consist of plotting all of the points (x, y, z) such
that f (x, y) = z. For example, here are a few surfaces:

p
Figure 3: Left to right: the hemisphere z = 1 − x2 − y 2 , parabola z = x2 + y 2 , and
monkey saddle z = x3 − 3xy 2 .

38
 Sometimes, we will want to visualize a surface even when we don’t have access to
computer programs! To do this, we use level curves, which are defined as follows: given a
function f (x, y), a level curve at height h is the set of all points (x, y) such that f (x, y) = h.
We think of this as what happens when we “slice” through the graph z = f (x, y) at height
z = h. If we take enough of these cross-sections, I claim that we get a nice visual image of
what our surface will look like!
For example, let f (x, y) = x2 − y 2 . I’ve drawn the level curves h = x2 − y 2 of this
function below, for values of h ranging from 4 to −4:
8 8 8

6 6 6

4 4 4

2 2 2

-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10

-2 -2 -2

-4 -4 -4

-6 -6 -6

-8 -8 -8

h=4 h=3 h=2

8 8 8

6 6 6

4 4 4

2 2 2

-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10

-2 -2 -2

-4 -4 -4

-6 -6 -6

-8 -8 -8

h=1 h=0 h=-1

8 8 8

6 6 6

4 4 4

2 2 2

-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10

-2 -2 -2

-4 -4 -4

-6 -6 -6

-8 -8 -8

h=-2 h=-3 h=-4

With some imagination, you can think about what it would look like if these level curves
were drawn in 3D space, each one at its corresponding height h, like this:

Indeed, if you fill in the gaps you can see the surface we’ve drawn here (a hyperbolic
paraboloid!)

39
14.2 Partial Derivatives
Given a function f (x, y), we define the partial derivative with respect to x of f (x, y),
d
denoted dx f (x, y), as the following: take f (x, y), think of x as a variable and y as a constant,
and take the derivative as normal with respect to x. So, for example,
d
(x + y) = 1 + 0 = 1,
dx
d d
sin(xy) = cos(xy) · (xy) = cos(xy) · y,
dx dx
d 2
y = 0.
dx
d
Similarly, we define the partial derivative with respect to y of f (x, y), denoted dy f (x, y),
as the following: take f (x, y), think of y as the variable and x as a constant, and take the
derivative with respect to y! So, for example,
d
(x + y) = 0 + 1 = 1,
dy
d d
sin(xy) = cos(xy) · (xy) = cos(xy) · x,
dy dy
d 2
y = 2y.
dy

Earlier, we used the derivative to make a tangent line to the graph y = f (x). We can
use the partial derivatives here to make a tangent plane to the graph z = f (x, y) at the
point (a, b, c) in a very similar way, using the equation below:
   
d  · (x − a) +  d (f (x, y))
z − c =  (f (x, y))  · (y − b)
dx dy
(x,y,z)=(a,b,c) (x,y,z)=(a,b,c)

We consider one last example to illustrate this idea:

Example. Find the tangent plane to the graph of f (x, y) = x2 + y 2 at the point (1, 1, 2).

Answer. We calculate:
d 2 d 2
(x + y 2 ) = 2x, (x + y 2 ) = 2y,
dx dy

Plugging this into the tangent plane formula yields

   
d  · (x − a) +  d (f (x, y))
z − 2 =  (f (x, y))  · (y − b)
dx dy
(x,y,z)=(a,b,c) (x,y,z)=(a,b,c)

⇒ z − 2 = 2(x − 1) + 2(y − 1).

Success!

40